Re: [Dynapi-Dev] TCanvas vs. DynLayer

[Richard B. <ma...@ri...>]

> You just don't get it.
Let's avoid this kind of disrespect where possible.

You've done some interesting work, and now that there's something tangible
to work with I spent some time testing it.
I can't comment on whether or not objects are being passed, or assigned
static or whatever, simply on test results.
I find the results show not necessarily how fast Canvas is, but how horribly
slow DynAPI has grown.
I did some extended testing, and put this online, so others can compare the
results for themselves.

I did the test without children, one building 1000 layers, one building 200
layers.
I applied these to:
* Canvas
* DynAPI 25/jan/01 release
* DynaCore 2001/01/13 release
* DynAPI Lite 29/aug/2000
* DynAPI Lite Compact code 29/aug/2000

My results where: (in ms. on p3 550 winMe IE5.5 128ram)
Canvas 1000            18.130
Canvas 200                1.320
DynAPI 1000           94.470
DynAPI 200                4390
DynaCore 1000        21.030
DynaCore 200            1.430
DynAPI Lite 1000       9.340
DynAPI Lite 200         1.200
DynAPI Comp.1000 11.590
DynAPI Comp. 200    1.100

The tests are online here:
http://www.dynapi.f2s.com/dynapi/tcanvas/

If anyone wants to add files or anything here's the FTP:
username: dynapi
password: dynapi

I think it's clear there is something seriously slowing down the DynAPI, if
DynaCore is so much faster with more or less the same functionality,
something's wrong.
Also I'm all for including the Lite files again, as their times are very
good.

Cheers,
Richard Bennett

ma...@ri...
www.richardinfo.com
(Everything running on, and ported to the 19/12/2000 snapshot of DynAPI2)
Find the DynAPI faq here:
http://sourceforge.net/docman/display_doc.php?docid=656&group_id=5757
Browse the mailinglist here:
http://www.mail-archive.com/index.php3?hunt=dynapi

----- Original Message -----
From: "Eytan Heidingsfeld" <ey...@tr...>
To: <dyn...@li...>
Sent: Sunday, February 18, 2001 9:57 AM
Subject: Re: [Dynapi-Dev] TCanvas vs. DynLayer


> You just don't get it.
> It's not that I use different code and the version I tested with has a
> children list. (uses my TList so didn't include it).
> The reason it's faster is such. When browsers parse js code and create
> objects they create actual objects in their application. That means that
for
> every dynlayer there is an object inside Netscape or IE with actual
objects
> in them. In your method of creation you:
> Create new DynLayer()
> Call add child to the parent
> Parent does the following
> static Create the virtual object
> static Create the actual element
> static Assign the element
> This means that the browser has to
> Get Parent object in browsers memory
> Call addchild
> Call static line and create js object
> Call static line and create js element
> Call staticassignment where you have to get obj1 and obj2 and assign
> and more of this type of thing
> It slows down the process by a whole lot
> What my create does is:
> Create new Canvas()
> myCanvas.create does this:
> Check who the parent is
> if null create in the actual document
> if Canvas create under it
> No passing of object between methods, not assigning no static methods.
>
> 8an
>
>
> _______________________________________________
> Dynapi-Dev mailing list
> Dyn...@li...
> http://lists.sourceforge.net/lists/listinfo/dynapi-dev
> ____________________________________________________________
> Get your free domain name and domain-based e-mail from
> Namezero.com. New!  Namezero Plus domains now available.
> Find out more at: http://www.namezero.com
>